arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.11204cs.AIcs.LG
收藏

复杂推理任务的子目标搜索

Subgoal Search For Complex Reasoning Tasks

University of Warsaw(华沙大学);University of Toronto(多伦多大学);Vector Institute(向量研究院);Polish Academy of Sciences(波兰科学院);University of Oxford(牛津大学)

Konrad Czechowski, Tomasz Odrzygóźdź, Marek Zbysiński, Michał Zawalski, Krzysztof Olejnik, Yuhuai Wu, Łukasz Kuciński, Piotr Miłoś

AI总结:

本文提出子目标搜索(kSubS)方法,利用Transformer生成可实现且接近解的子目标,结合best-first搜索,在Sokoban、魔方和INT基准上取得高效表现,其中INT达到最先进水平。

摘要:

人类擅长通过从一个想法转移到相关想法的心理过程来解决复杂的推理任务。受此启发,我们提出了子目标搜索(kSubS)方法。其关键组件是一个学习得到的子目标生成器,它产生多样化的子目标,这些子目标既可实现又更接近解决方案。使用子目标减少了搜索空间,并引出了一个适合高效规划的高层搜索图。在本文中,我们使用基于Transformer的子目标模块结合经典的best-first搜索框架实现了kSubS。我们展示了生成第$k$步提前子目标的简单方法在三个具有挑战性的领域上出奇地高效:两个流行的益智游戏,Sokoban和Rubik's Cube,以及一个不等式证明基准INT。kSubS在适度的计算预算内取得了包括在INT上达到最先进水平在内的强劲结果。

英文摘要:

Humans excel in solving complex reasoning tasks through a mental process of moving from one idea to a related one. Inspired by this, we propose Subgoal Search (kSubS) method. Its key component is a learned subgoal generator that produces a diversity of subgoals that are both achievable and closer to the solution. Using subgoals reduces the search space and induces a high-level search graph suitable for efficient planning. In this paper, we implement kSubS using a transformer-based subgoal module coupled with the classical best-first search framework. We show that a simple approach of generating $k$-th step ahead subgoals is surprisingly efficient on three challenging domains: two popular puzzle games, Sokoban and the Rubik's Cube, and an inequality proving benchmark INT. kSubS achieves strong results including state-of-the-art on INT within a modest computational budget.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑